Author Topic: Compound Noun Detection  (Read 42 times)

Offline bballboy8

  • New Linguist
  • *
  • Posts: 1
Compound Noun Detection
« on: April 01, 2022, 10:57:27 AM »
Hey everyone!

I work doing natural language processing in machine learning and I need to be able to detect compound nouns in text. Unfortunately, I haven't been able to find a comprehensive list online. Would such a list exist and if so, where would I find it? If not, is there some way to systematically detect compound nouns in text?

Offline Daniel

  • Administrator
  • Experienced Linguist
  • *****
  • Posts: 2090
    • English
Re: Compound Noun Detection
« Reply #1 on: April 01, 2022, 12:35:48 PM »
Practical answer: a list might work to identify (some) compound nouns, but I don't know of one to suggest.

But a list would not include all compound nouns because we can productively create them through usage, such as "noun list".

Theoretically, there's also a question about how to define compound nouns: are these two nouns in a row in text? What about one-word compounds (e.g. "birdhouse")? Yes, you might need a list for those, but it would be tricky because some people might make up new ones (like "nounlist") spontaneously, but a list should cover at least the most common ones written as one word. That, plus a list of nouns (or just part of speech tagging) that would identify two nouns in a row would find a lot more. (And hyphenated compounds.) But however you approach this you might have some false positives and false negatives. Morphologists don't have an exact definition of "compound noun" that would always work perfectly, especially for automatic text processing. The best you can do, if you want an "easy" (and practical) answer, is to choose relatively simple criteria that will catch most cases. Or you can use a list, but I don't see how much insight that would give unless it is a very comprehensive list.
Welcome to Linguist Forum! If you have any questions, please ask.